at SciVerse ScienceDirect

Polymer 53 (2012) 2964e2972

Contents lists available

Polymer

journal homepage: www.elsevier .com/locate/polymer

Phthalimide based polymers of intrinsic microporosity

Saad Makhseed*, Fadi Ibrahim, Jacob SamuelDepartment of Chemistry, Kuwait University, P.O. Box 5969, Safat 13060, Kuwait

a r t i c l e i n f o

Article history:Received 29 December 2011Received in revised form18 April 2012Accepted 2 May 2012Available online 10 May 2012

Keywords:Microporous polymersMembraneMolecular weight distribution

* Corresponding author. Tel.: þ965 24985538; fax:E-mail address: saad.makhseed@ku.edu.kw (S. Ma

0032-3861/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.polymer.2012.05.001

a b s t r a c t

A series of phthalimide based microporous polymers were successfully prepared by conventionalnucleophilic substitution reaction of several newly synthesized fluoro-monomers with commerciallyavailable 5,50 ,6,60-tetrahydroxy-3,3,30,30-tetramethylspirobisindane. FTIR, 1H NMR, and elemental anal-yses were used to identify the proposed structures of the polymers. The synthesized polymers are of highmolecular weight as demonstrated by Gel Permeation Chromatography (GPC). Thermogravimetricanalysis shows that the prepared polymers were stable up to 300 �C. From the porosity analysis it is clearthat the prepared polymers are analogous to polymers of intrinsic microporosity (PIMs) with highsurface area (500e900 m2/g). The t-plot analysis shown that the major contribution to the specificsurface area is arising from the micropore surface area with narrow size distribution of ultramicroporesas confirmed by the Horvath-Kawazoe (H-K) analysis. The hydrogen storage capacity of the preparedPIM-R(1e7) and CO-PIM(3,4,6,7) were promising (up to 1.26 wt%, 77 K, at 1.13 bar) with high isotericheats of H2 adsorption (8.5 kJ/mol).The results of this study demonstrate that controlling the appropriatemonomer content via the three-dimensional structure can provide a uniform microporous morphologyin the target polymers.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The preparation of purely organic microporous polymerswithout the assistance of molecular template is becoming a fastdeveloping area in nanomaterial’s research and many approacheshave been used to develop various insoluble network organicmicroporous polymers such as hyper cross-linked polymers (HCP)[1,2], triptycene-based PIM (Trip-PIM) [3], organic frameworkpolymer (OFP) [4], covalent triazine frameworks (CTFs) [5] andmicroporous polyimide networks [6] with high specific surface.Conjugated microporous polymers (CMPs) and their analogues areanother important class of insoluble microporous organic polymerswhich can be considered as cross-linked polymers with extendedconjugated network [7]. On the other hand, the inclusion of certainfunctionalities into CMPs can affect their surface areas incomparison with their non-functionalized analogues [8]. Recentlymany reviews published on nanoporous polymers and theirpotential applications with a focus on microporosity [9e12]. Themicroporosity is maintained by a robust network of covalent bonds,further complemented by rigid framework structure. Generallynon-network polymers can pack space efficiently leading to

þ965 24816482.khseed).

All rights reserved.

nonporous material. It is important to note that if there is a certainamount of free volume, voids would be interconnected, and thepolymer will therefore behave as a microporous material evenwithout a network structure. This kind of microporous polymercould be soluble to facilitate easy solution-based processing whichcan’t be achieved in other microporous materials. As a result theyhave attracted good acceptance as outstanding membrane in gasseparation, gas storage, adsorption of small molecules andheterogeneous catalysis [13]. So it is of great interest to synthesizehighly soluble porous polymers using the established dioxane-forming polymerization reaction between appropriate fluorine-containing monomers and the commercially available aromatictetrol. The porous properties mainly due to the architecturalfeatures of the used monomers which defines the microporosity.The combined features: high rigidity arising from dioxane linkagesand the randomly contorted structure due to either a spiro-centreor a non-planar architecture, represent a prerequisite to inducespace inefficient packing producing a large amount of inter-connected free volume. The most outstanding representation forthis kind of materials are the polymers of intrinsic microporosity(PIMs) [14] developed by McKeown et al with high surfaceareas in the range of 500e900 m2/g. Typical example is PIM-1,prepared as a soluble polymer by the dioxane-forming reactionbetween 5,50,6,60-tetrahydroxy-3,3,30,30-tetramethylspirobisindaneand tetrafluoroterephthalonitrile. The same concept of using rigid

S. Makhseed et al. / Polymer 53 (2012) 2964e2972 2965

and space inefficient packing subunits was expanded by Guiveret al. in developing a high molecular weight PIM-1; synthesized bypolycondensation under high-intensity mixing conditions at about155 �C to yield cyclic-free products in a shorter time (8 min) [15].McKeown et al, additionally synthesized a highly soluble micro-porous polymer derived from Bis(phenazyl) monomers withreasonable internal surface area and good film-forming character-istics (PIM-7) [16]. Studies on evaluating the influence of inter-molecular interactions in microporous polymers shown that suchinteractions (e.g. hydrogen bonding) can lead to hindered orreduced microporosity [17,18]. The structural features of PIM-1further tailored by post- polymerization modifications throughthe conversion of nitrile moieties into different functional groupssuch as tetra azole and thioamide with enhanced CO2 capture andgas transport properties [19,20]. To date, several soluble ladderpolymers such as polybenzodioxanes, polyimides, and PIMs con-taining dinaphtyl or tianthrene segments have been reported andthe influence of their modified chemical structure on the physicalproperties, sorption behaviour and gas permeability has beenexplored [21e23]. Although the structural alteration of PIM-1 led tolow surface areas, these materials showed promising results forcertain applications [24].

Controlling the pore diameter at the micro scale (IUPAC pores ofdimension < 2 nm) make it possible to use such materials indifferent commercial applications such as hydrogen adsorption[13]. The lack of fundamental understanding of interaction betweenthe polymer surfaces and the gas molecules initiate manyapproaches with the aim of designing polymer materials to estab-lish the structure-porous property relationship. This may be ach-ieved by selecting appropriate monomer precursors with certainstructural features. For example, monomers decorated by largersubstituents can frustrate the resulting polymer chain packingwhich may create higher surface area along with tunable porosity.In the present study we described a series of solution processablephthalimide based organic polymers prepared from inexpensivemonomers with sufficient surface area and narrow pore diameterdistribution to store H2 in the confined space through enhancedinteraction. While considering hydrogen storage application theprepared microporous polymers (PIM-Rs and CO-PIM-Rs) offermany attractive features such as low intrinsic density, structuralhomogeneity, hydrothermal stability and synthetic reproducibility.The chemical features of the surfaces and nature of the pores havethus tuned to achieve a high hydrogen adsorption.

2. Experimental

2.1. Characterization

1H- NMR spectrum (400 MHz) of monomers and polymers wererecorded on a Bruker DPX 400 spectrometer using CDCl3 as thesolvent and tetramethylsilane as the internal standard. FTIR spectrawere recorded on a JASCO FT/IR-6300. Elemental analyses werecarried out using Elementar Vario Micro Cube. Mass analyses weredone on a Thermo DFS Mass spectrometer. The molecular weightand its distribution were measured by Knauer Gel PermeationChromatograph fitted with a refractive index detector (flow rate1 mL/min) according to polystyrene standards using CHCl3 as theeluent. Thermogravimetric analysis (TGA) was performed witha Shimadzu TGA-50 instrument at a heating rate of 10 �C/min undernitrogen atmosphere and DSC analysis was done on Shimadzu DSC-60 instrument at a heating rate of 10 �C/min under nitrogen.Melting points were measured with a Griffin melting point appa-ratus and further confirmed by DSC analysis. Wide-angle X-raydiffraction (WAXD) of films was measured by a Siemens D5000diffractometer. Microscopic techniques employed are Scanning

Electron Microscopy (SEM: JEOL Model 6300) and High ResolutionTransmission Electron Microscopy (HRTEM: JEOL Model JEM-3010). Nitrogen adsorption measurements (77 K) and volumetrichydrogen adsorptions analysis (77 K and 87 K at 1.13 bar) wereperformed on a Micromeritics ASAP 2020 Sorptometer equippedwith an out gassing platform, an online data acquisition andhandling system. Before analysis the samples were degassed for12 h with a heating rate of 1 �C/min in two stages (80 �C for 1 h and120 �C for 11 h) under high vacuum (<10�4 mbar). The specificsurface area was calculated using the Brunauer-Emmet-Teller (BET)equation (See supporting Information). The micropore area wascalculated using t-plot method (Harkins and Jura). The pore sizedistributions were calculated from the adsorption isotherm usingthe Horvath-Kawazoe (H-K) calculations (See supporting Informa-tion). The heats of adsorption were calculated from the hydrogenadsorption isotherms obtained at 77.3 K and 87 K using the ASAP2020 software (Micromeritics, Norcross, GA).

2.2. Materials

The anhydrous solvents DMF and DMAc were purchased fromAldrich co, and the potassium carbonate obtained from Merck wasfinely grounded and dried at 200 �C. Tetrafluoroterephthalonitrileobtained from aldrich was used directly. 5,50,6,60-tetrahydroxy-3,3,30,30-tetramethyl-1,10-spirobisindane obtained from alfa waspurified from methanol. Starting from tetrafluorophthalic acid,tetrafluorophthalic anhydride was prepared in our laboratory. Theother commercially available materials and solvents were usedwithout further purification.

2.3. Film preparation

Polymer (550 mg) was dissolved in CHCl3 (15 mL) and filtered(2 mm glass microfiber). The filtered solutionwas pouring into a flatPetri dish (12 cm dia.) and then placed in a vacuum desiccator withthe gas inlet open to the atmosphere. The solution was allowed toevaporate for 4 days. The dish was removed and the clear yellowfilm released from the glass.

2.4. Synthesis of fluoro-monomers

2.4.1. 4,5,6,7-Tetrafluoro-2-phenyl-isoindole-1,3-dione. MR1Tetrafluorophthalic anhydride (8.5g, 38.7 mmol) was added to

a stirred solution of aniline (3.00 g, 32.26 mmol) in glacial aceticacid and refluxed for 12 h. The white solid obtained was furtherrefluxed in acetic anhydride for 12 h. After cooling to roomtemperature, the precipitated product was filtered off and washedwith petroleum ether to give a light yellow solid. Yield: 84%; Mp

206 �C. 1H NMR (CDCl3, d ppm): 7.26-7.54 (m, 4H, Ar). FTIR(KBr pellet, cm�1): 3068, 1783 (asymmetry C]O), 1726 (symmetryC]O), 1460 (C]C), 1371 (C-N), 945 (C-F). Anal. Calcd forC14H5NO2F4 (295.08): C, 56.94; H, 4.74; N, 4.74. Found: C, 56,88; H,1.80; N, 5.03. MS (EI): m/z (%) 295 (100%).

The following monomers were prepared from tetra-fluorophthalic anhydride and corresponding amines using similarprocedures adopted for MR1.

2.4.2. 4,5,6,7-Tetrafluoro-2-hexyl-isoindole-1,3-dione. M-R2Yield: 88%; Mp 118 �C. 1H NMR (CDCl3, d ppm): 3.68 (t, 2H, CH2),

1.63-1.70 (m, 2H, CH2), 1.29-1.35 (m, 6H, CH2), 0.9 (t, 3H, CH3). FTIR(KBr pellet, cm�1): 3078, 1780 (asymmetry C]O), 1720 (symmetryC]O), 1465 (C]C), 1370 (C-N), 944 (C-F). Anal. Calcd forC14H13NO2F4 (303.26): C, 55.45; H, 4.32; N, 4.62. Found: C, 54.70; H,4.01; N, 4.84. MS (EI): m/z (%) 303 (100%).

S. Makhseed et al. / Polymer 53 (2012) 2964e29722966

2.4.3. 4,5,6,7-Tetrafluoro-2-(4-methoxy-phenyl)-isoindole-1,3-dione. M-R3

Yield: 85%; Mp 250 �C. 1H NMR (CDCl3, d ppm): 7.28- 7.24 (dd,2H, Ar), 7.06-7.02 (dd, 2H, Ar), 3.85 (s, 3H, CH3). FTIR (KBr pellet,cm�1): 3070, 2930,1775 (asymmetry C]O),1730 (symmetry C]O),1468 (C]C), 1338 (C-N), 952 (C-F). Anal. Calcd for C15H7 NO3F4(325.03): C, 55.40; H, 2.17; N, 4.31. Found: C, 55.50; H, 2.07; N, 4.43.MS (EI): m/z (%) 325 (100%).

2.4.4. 4,5,6,7-Tetrafluoro-2-trimethoxybenzene-isoindole-1,3-dione. M-R4

Yield: 87%; Mp 217 �C. 1H NMR (CDCl3, d ppm): 6.61 (s, 2H, Ar),3.91 (s, 3H, CH3), 3.88 (s, 6H, CH3). FTIR (KBr pellet, cm�1): 3087,2935, 1760 (asymmetry C]O), 1728 (symmetry C]O), 1462 (C]C),1350 (C-N), 960 (C-F). Anal. Calcd for C17H11NO5F4 (385.10):C, 52.98; H, 2.89; N, 3.64. Found: C, 52.58; H, 2.88; N, 3.85. MS (EI):m/z (%) 385 (100%).

2.4.5. 4,5,6,7-Tetrafluoro-2- 4-tert -butylbenzene e isoindole-1,3-dione. M-R5

Yield: 84%; Mp 221 �C. 1H NMR (CDCl3, d ppm): 7.54e7.51 (dd,2H, Ar), 7.30e7.28 (dd, 2H, Ar), 1.35 (s, 9H, CH3). FTIR (KBr pellet,cm�1): 3095, 2950,1772 (asymmetry C]O),1735 (symmetry C]O),1458 (C]C), 1350 (C-N), 960 (C-F). Anal. Calcd for C18H13NO2F4(351.1): C, 61.52; H, 3.73; N, 3.98. Found: C, 61.70; H, 3.49; N, 4.17.MS (EI): m/z (%) 351 (45%).

2.4.6. 4,5,6,7-Tetrafluoro-2-(2,6-diisopropyl-phenyl)-isoindole-1,3-dione. M-R6

Yield: 92%; Mp 171 �C. 1H NMR (CDCl3, d ppm): 7.48 (t, H, Ar),7.30-7.28 (d, 2H, Ar), 2.63 (sept, 2H, CH), 1.17-1.16 (d,12H, CH3). FTIR(KBr pellet, cm�1): 3090, 2940, 1770 (asymmetry C]O), 1733(symmetry C]O), 1450 (C]C), 1350 (C-N) and 970 (C-F). Anal.Calcd for C20H17NO2F4 (379.10): C, 63.30; H, 4.48; N, 3.69. Found: C,63.12; H, 4.26; N, 3.85. MS (EI): m/z (%) 379 (50%).

2.4.7. 4,5,6,7-Tetrafluoro -1-adamantylamine. M-R7Yield: 73%; Mp > 300 �C. 1H NMR (CDCl3, d ppm): 2.44 (s, 6H,

CH2), 2.15 (s, 3H, CH), 1.77-1.74 (d, 3H, CH2), 1.71-1.68 (d, 3H, CH2).FTIR (KBr pellet, cm�1): 3103, 2911, 1779 (asymmetry C]O), and1732 (symmetry C]O), 943 (C-F). Anal. Calcd for C18H15NO2F4(353.10): C, 61.17; H, 4.24; N, 3.96. Found C, 60.86; H, 4.40; N, 4.08.MS (EI): m/z (%) 353 (100%).

2.5. Synthesis of PIM-R(1e7)

2.5.1. PIM-R1A mixture of monomer MR1 (0.50 g, 1.69 mmol) and spi-

robiscatechol (0.57 g, 1.69 mmol) in dry DMF (50 ml) and K2CO3(0.50g, 3.51 mmol) was stirred at 120 �C for 4 h. After cooling, thereaction mixture was poured into deionised water and the solidproduct collected byfiltrationwashedwithmethanol. The insolublepolymer was then purified by refluxing in methanol and acetone.The obtained yellow fluorescent powder was dried in vacuum ovenat 100 �C for 12 h. Yield: 75%.Mp> 300 �C.Mw 42,403.Mn 40,152. 1HNMR (CDCl3, 298 K d ppm): 7.68-7.38 (s, br, 3H, Ar), 6.87-6.51 (d, br,4H, Ar), 2.30 (d, br, 4H, CH2), 1.30 (s, br, 12H, CH3). FTIR (KBr pellet,cm�1): 3088, 1774 and 1723 (imide). Anal. Calcd for C35H25NO6

(555.45): C, 75.61; H, 4.50; N, 2.52. Found: C, 75.19; H, 4.43; N, 2.45.The following polymers were prepared from corresponding

fluoro-monomers using similar procedure adopted for PIM-R1.

2.5.2. PIM-R2Yield: 82%. Mp > 300 �C. Mw 27,624. Mn 15,741. 1H NMR (CDCl3,

298 K, d ppm): 6.91(s, br, Ar), 6.50 (s, br, Ar), 3.45 (m, br, 2H, CH2),

2.18 (d, br, 2H, CH2), 1.58-1.30 (d, br, 12H, CH3). FTIR (KBr pellet,cm�1): 3080, 2940, 1774 and 1709 (imide). Anal. Calcd forC35H33NO6 (563.01): C, 74.60; H, 5.86; N, 2.04. Found: C, 74.28; H,5.77; N, 2.01.

2.5.3. PIM-R3Yield: 84%.Mp > 300 �C.Mw 178,843.Mn 62,276. 1H NMR (CDCl3,

298 K, d ppm): 6.95-6.77 (d br, Ar), 6.5 (s br, Ar), 3.85 (t br, OCH3),2.30-2.17 (d, br, CH2), 2.17 (d, 2H, CH2), 1.31 (s, 12H, CH3). FTIR (KBrpellet, cm�1): 3090, 1768 and 1717 (imide). Anal. Calcd forC36NO7H27 (585.07): C, 73.84; H, 4.61; N, 2.39. Found: C, 73.48; H,4.52; N, 2.26.

2.5.4. PIM-R4Yield: 84%.Mp > 300 �C.Mw 214,278.Mn 69,348. 1H NMR (CDCl3,

298 K, d ppm): 6.89 (s, 2H, Ar), 6.91-6.53 (m br, Ar), 3.90-3.79 (m br,OCH3), 2.33-2.18 (d, CH2), 1.72 (s, 2H, CH2), 1.51(s, 12H, CH3). FTIR(KBr pellet, cm�1): 1779 and 1712 (imide). Anal. Calcd forC38H31NO9 (645.03): C, 70.69; H, 4.80; N, 2.17. Found: C, 70.57; H,4.75; N, 2.13.). Anal. Calcd for C38H31NO9 (645.03): C, 70.69; H, 4.80;N, 2.17. Found: C, 70.37; H, 4.71; N, 2.13.

2.5.5. PIM-R5Yield: 86%; Mp > 300 �C. Mw 37,200. Mn 31,000. 1H NMR (CDCl3,

298 K, d ppm): 7.44-7.31 (d br, Ar), 6.88-6.51 (d br, Ar), 2.30-2.17(d br, 4H, CH2), 1.29-1.25 (d, 21H, CH3). 3086, 2954, 1771 and 1723(imide). Anal. Calcd for C39H33NO6 (611.24): C, 76.59; H, 5.40; N,2.29. Found: C, 76.28; H, 5.32; N, 2.21.

2.5.6. PIM-R6Yield: 87%.Mp > 300 �C.Mw 162,594.Mn 67,748. 1H NMR (CDCl3,

298 K, d ppm): 7.43 (d, 2H, Ar), 6.85 (t br, 2H, Ar), 6.53 (s br, Ar)2.70(s br, 2H, CH) 2.34-2.09 (d br, 4H, CH2) 1.32-1.17 (d, 24H, CH3).FTIR (KBr pellet, cm�1): 3088, 2945, 1781 and 1727 (imide). Anal.Calcd for C41H37NO6 (641.05): C, 76.90; H, 5.75; N, 2.21. Found: C,76.79; H, 5.68; N, 2.17.

2.5.7. PIM-R7Yield: 80%.Mp > 300 �C.Mw 152,916.Mn 27,803. 1H NMR (CDCl3,

298 K, d ppm): 6.84-6.72 (d br, 2H, Ar), 6.46 (s br, 2H, Ar), 2.51-2.39(t br, CH), 2.17-210(d br, 4H CH2), 1.71-1.67(d br,12H, CH2),1.29 (s br,12H, CH3). FTIR (KBr pellet, cm�1): 3098, 2960, 1781 and 1709(imide). Anal. Calcd for C39H35NO6 (613.08): C, 76.30; H, 5.70; N,2.28. Found: C, 76.08; H, 5.61; N, 2.19.

2.6. Synthesis of CO-PIM-R(3,4,6,7)

2.6.1. CO- PIM-R3Amixture ofmonomerMR3 (0.20g, 0.52mmol), spirobiscatechol

(0.35g, 1.04 mmol) and 2,3,5,6-tetrafluorophthalonitrile (0.10g,0.52 mmol) in dry DMAc (40 ml) and K2CO3 (0.36g, 2.62 mmol) wasstirred at 120 �C for 4 h. After cooling, the reaction mixture waspoured into deionised water and the solid product collected byfiltration and washed with methanol. The insoluble polymer wasthen purified by refluxing in methanol, acetone and THF. Theobtained yellow fluorescent powder was dried in vacuum oven at100 �C for 12 h. Yield: 85%.Mp> 300 �C.Mw 155,908.Mn 125,753. 1HNMR (CDCl3, 298 K, d ppm): 6.95-6.89 (d br, 2H, Ar), 6.51 (s, 2H, Ar),3.82 (t br, 3H, OCH3), 2.30-2.17 (d, 4H, CH2), 1.31 (s, 12H, CH3). FTIR(KBr pellet, cm�1): 3070, 2236 (C^N), 1768 and 1720 imide). Anal.Calcd for C44H37N3O9 (741.19): C, 71.25; H, 3.67; N, 5.67. Found: C,70.93; H, 3.52; N, 5.55.

Following copolymers were prepared from correspondingfluoro-monomers using similar procedure adopted for CO- PIM-R3.

S. Makhseed et al. / Polymer 53 (2012) 2964e2972 2967

2.6.2. CO- PIM-R4Yield: 80%. Mp > 300 �C. Mw 186,930. Mn 148,563. 1H NMR

(CDCl3, 298 K, d ppm): 7.31e7.27 (dd, 2H, Ar), 7.10-7.06 (dd, 2H, Ar),3.85 (s, 3H, CH3). FTIR (KBr pellet, cm�1): 3088, 2239(C^N), 1767and 1721 (imide). Anal. Calcd for C46H31N3O11 (801.23): C, 68.91; H,3.90; N, 5.24. Found: C, 68.48; H, 3.74; N, 5.16.

2.6.3. CO-PIM-R6Yield: 83%. Mp > 300 �C. Mw 162,000. Mn 142,000. 1H NMR

(CDCl3, 298 K, d ppm): 7.43e7.28 (d br, 2H, Ar) 6.94-6.90 (d br, 2H,Ar), 6.52-6.44(d br, 2H, Ar), 3.96 (s, 3H, CH3), 3.80 (s, 6H, CH3). FTIR(KBr pellet, cm�1): 3090, 2236 (C^N), 1768, 1721 (imide). Anal.Calcd for C49H37N3O8 (795.21): C, 73.95; H, 4.69; N, 5.28. Found: C,73.55; H, 4.41; N, 5.19.

2.6.4. CO- PIM-R7Yield: 83%. Mp > 300 �C. Mw 178,808. Mn 139,788. 1H NMR

(CDCl3, 298 K, d ppm): 6.87 (d br, 2H, Ar), 6.48 (s br, 2H, Ar), 2.54-2.42 (t br, CH), 2.25-2.17(d br, 4H CH2), 1.73-1.69(d br, 12H, CH2),1.32 (s br, 12H, CH3). FTIR (KBr pellet, cm�1): 3111, 2965, 2236(C^N), 1777 and 1731 (imide), Anal. Calcd for C47H35N3O8 (769.82):C, 73.33; H, 4.58; N, 5.46. Found C, 73.15; H, 4.42; N, 5.34.

3. Results and discussion

The phthalimide based fluorinated monomers (MR [1-7]) wereprepared in good yield by the straight forward one step imidisationreaction between tetrafluorophthalic anhydride and differentamines in refluxing acetic acid. The proposed structure and purity ofthe obtained fluorinated monomers were confirmed by routinelyspectroscopic techniques as well as elemental analysis. The possi-bility of the preparation of ladder polymerwas confirmed bymakingmodel compounds directly from a reaction between 2 M equivalentsof di-tertiarybutyl catechol with the different tetraflurophthalimide(See Supporting Information). The reaction product was character-ized by mass, 1H NMR, FTIR spectroscopy and elemental analysisconfirming that all the four F atoms were replaced by the two ben-zodioxane units. The highly efficient reaction (yield>90%) promotesthe preparation of microporous organic polymers. The polymers(PIM-R[1e7]) were synthesized by the dibenzodioxane formationreaction between the 5,50,6,60-tetrahydroxy-3,3,30,30-tetramethyl-spirobisindane (spirobiscatechol), and various fluorine-containingmonomers (MR[1e7]) in dry DMF as illustrated in Scheme 1. Thepolymers were isolated by precipitation in acidified water and theresulting powder samples were purified by refluxing in deionised

N OO

F

FF

F

R

OH

OH

OH

OH

OMe

MeO

+

R1 = R2 = R3 =

R5 = R7 =R6 =

R4 =

(1-7)

Scheme 1. Synthesis of microporous polymers PIM-R(1e7)s. Reag

water, methanol and ethanol respectively followed by reprecipi-tation in a chloroform/methanol mixture to give yellow fluorescentpowder in good yield (>85%) attributed to the efficiency of thepolycondensation reaction. The solution processable PIM-Rs can becast into membrane which can be used as a separation membrane,selectively removing one component either from liquid or gasmixture. According to our previous experiments, we note that thesynthesis of PIMs having high molecular weight could be carriedout at revised conditions especially elevated temperature and highconcentration [25]. After a number of trials, we found that thesolvent DMF is significantly compatible with the reaction mixtureto a great extent and could achieve high molecular weight linearpolymers while compared to DMAc and NMP. We optimized thereaction conditions to get high molecular weight polymers usingDMF as a solvent that is significantly well-suitedwith themonomersalts and the polymers dissolved in the reaction mixture at highconcentration conditions. The average molecular masses ofprepared polymers were estimated by gel permeation chromatog-raphy (GPC) relative to polystyrene standard in CHCl3 (Fig. 1). Thedata obtained from GPC is listed in Table 1. PIM-R3, PIM-R4 andPIM-R6 with relatively high average molecular weight wereobtained. The high molecular weight can be attributed to the goodsolubility of randomly forming polymers in the reaction mixture.On the other hand, GPC curves (See supporting Information) showshoulder peaks in the low molecular weight region while PIM-R5with reasonable molecular weight was obtained in a quantitativeyield without any shoulder peak. This behaviour can be attributedto the presence of cyclic oligomers as a result of the side reactions. Ithas been reported that in classic polycondensation reactions ofbisphenols and bifluoro aromatics for the production of poly(arylethers) that a large amount of macrocyclic oligomers were formed[26,27]. During the first step of the base catalysed reaction, the tetrafunctional spirobiscatechol reacted with a base to give a number ofsalt which have low solubility and resulted in the formation ofcyclic compounds. The lowmolecular weight, in the case of PIM-R2,probably caused by either less soluble salts of propagating chains orcyclic oligomer formation. However the reaction rate whichdepends on the structural features and the reactivity of eachmonomer also represent another key factor in determining themolecular weight. This clearly indicates that the rate-controllingstep in the reaction is the dissolution of the salt. Using thesimilar procedure adopted for PIM-Rs a series of new copolymers(CO-PIM-R3, CO-PIM-R4, CO-PIM-R6 and CO-PIM-R7) by employ-ing tetrafluoroterephthalonitrile were prepared. After many trialprocedures by using different feed ratios of monomers, the

O

O

O

O

NO

RO

OO

OMeOMe

n

PIM-R(1-7)

(1-7)

ents and conditions: Dry DMF, Anhydrous K2CO3, 120 �C, 4 h.

N OO

F

FF

F

R

OH

OH

OH

OH

O

O

NO

RO

OOO

O

CN

CN

F

F F

FCN

CN

+

(3,4,6,7)

nCO-PIM-R(3,4,6,7)

(3,4,6,7)

+

Scheme 2. Synthesis of microporous polymers CO-PIM-R(3,4,6,7). Reagents andconditions: Dry DMF, Anhydrous K2CO3, 120 �C, 4 h.

Fig. 1. GPC curves of PIM-R5 and CO-PIM-R6 with narrow molecular weightdistribution.

S. Makhseed et al. / Polymer 53 (2012) 2964e29722968

optimum ratio of fluorine-containing monomers and 2,3,5,6-tet-rafluorophthalonitrile, which has been applied to produce themaximum molecular weight copolymer is 0.5:0.5. with highmolecular weights (Mn; 125,753e148,563). Although the reaction iscomplicated due to the monomers multiple reactive groups CO-PIM-Rs show GPC curves equivalent to a linear unbranched chainfree of macrocyclic species and crosslinking (See SupportingInformation). Thus, under the same reaction conditions molecularweight broadening and crosslinking are efficiently reduced byintroducing a fixed ratio of tetrafluorophthalonitrile into the reac-tion mixture, and high-molecular-weight copolymers can beobtained. It is understood that in the reaction mixture tetra-fluorophthalonitrile has a higher reactivity, and its polar groupscreate them active in the reaction system, by the formation ofrandomly forming the copolymer to a great extent which results inless crosslinking and formation of cyclic oligomers. On the otherhand, by using chlorine containing structurally equivalent mono-mers at different conditions, the high-molecular-weight copoly-mers still could not be obtained (Scheme 2).

The structures of all PIM-Rs and CO-PIM-Rs were characterizedby FTIR, 1H NMR spectroscopy and elemental analysis. All PIMsretained its characteristic stretching bands related to the imidegroups (C]O symmetric and asymmetric stretching in the range of1722e1792). Moreover it is clear from the spectrum that the basecatalysed polymerization condition does not make any destructionin the imide units (See Supporting Information). In the 1H NMRspectrum signal positions of PIM-Rs and CO-PIM-Rs were located

Table 1GPC results and mechanical properties of PIM-Rs and CO-PIM-Rs and theircomparison with PIM-1.

PIMs Mn Mw Mw/Mn Tensilestrength (MPa)

PIM-1 [14b,23] 96,428 270,000 2.8 47.1PIM-R1 40,152 42,403 1.27 48.1PIM-R2 15,741 27,624 1.75 e

PIM-R3 62,276 178,843 2.87 45.0PIM-R4 69,348 214,278 3.09 46.5PIM-R5 31,000 37,200 1.20 39.4PIM-R6 67,748 162,594 2.38 48.6PIM-R7 27,803 152,916 5.50 41.0CO-PIM-R3 125,753 155,908 1.24 49.0CO-PIM-R4 148,563 186,930 1.25 48.4CO-PIM-R6 142,000 162,000 1.1 48.8CO-PIM-R7 139,788 178,808 1.27 44.0

precisely with the help of 1H NMR spectra of monomers andprecursor molecules. The 1H NMR spectra of the PIMs were fullyconsistent with their proposed structures (See SupportingInformation). The signals originating from the hydroxyl groups ofspirobiscatechol disappeared, suggesting that the reaction wascarried out completely through dibenzodioxane formation. Thebroad aromatic and aliphatic signals of PIMs were enough toconfirm the formation of high molecular weight polymers. Thethermal properties were evaluated by TGA and DSC. The decom-position temperatures were also very high; where a 5e10% initialloss around 300 �C, were observed, corresponding to the vapor-ation of the residual solvents which has been identified as theentrapped solvents in the micropores used for processing thesample (See Supporting Information). The good thermal stabilitycan be attributed to its double stranded structures. In DSC analysisno melting (Tm) and glass transition (Tg) was observed. Wide AngleX-ray Diffraction (WAXD) analysis of the films were conducted todisplay no crystalline peaks and revealed that all the preparedmaterials were amorphous. The solubility of the PIM-Rs and CO-PIM-Rs in various organic solvents were tested and found to behighly soluble in CHCl3, and partially in THF, and DMF. The PIMsshowed good flexibility, transparency, and the strong thin filmscould be obtained readily by casting from CHCl3 solutions in allcases (Fig. 2). The mechanical properties of these samples weretested using dumb-bell shaped specimens, except PIM-R2, whichdid not form a good membrane. As shown in Table 1, tensilestrength of all the prepared polymers was comparable irrespective

Fig. 2. Transparent free standing PIM-R6 film (Mn ¼ 67748, PDI ¼ 2.38, thickness70 mm).

Fig. 3. SEM images of CO-PIM-R6 a) powder and b) thin film.

S. Makhseed et al. / Polymer 53 (2012) 2964e2972 2969

of their molecular weight. This is attributed to the good structuralrigidity due to the presence of strong double stranded backbone(ladder-like sequences). As usually the mechanical properties oflinear polymers depend on its molecular weight.

The particle size and morphology of microporous polymers wereevaluated by scanning electronmicroscopy (SEM). Fig. 3 depicts SEMimages of the powder and film samples of CO-PIM-R6. The phasehomogeneity and uniform size distribution of particles is observed inmost of the precipitated samples with considerable differences insurface morphologies of the prepared polymers. The SEM images ofthe precipitated samples also revealed an aggregation of looselypacked spherically as well as non-spherically shaped micro andnanoparticles (See Supporting Information). The formation of lessporous structure in the film can be ascribed to the more denselypacked of polymer chains after the solvent evaporation. However thedifference in the polymer microstructure due to processing is sosignificant while considering the sorption ability. Using transmissionmicroscope at high resolution (HRTEM), polymers are seen to haveporous structure (Fig. 4). Overall, this is a fairly unique nanoporousstructure that dramatically enhances adsorption of small molecules.Like other porous materials (e.g. activated carbons) our preparedpolymers were found to be stable under the HRTEM analysis(300 kV) as an additional indication of the high structural stability.

3.1. N2 sorption analysis

The porous nature of PIMs was analysed by means of nitrogensorption at 77 K. The most common technique used for screeningthe porosity involves the fitting of Brunauer-Emmett-Teller (BET)

Fig. 4. HRTEM images of PIM-R4 pow

equation to the data obtained from the nitrogen adsorption/desorption isotherms at 77 K. Table 2 lists several relevantparameters obtained from the isotherm analysis including surfaceareas, micropore area, micropore volume and pore diameter basedon Horvath-Kawazoe (H-K). The BET surface areas of PIM-R(1e7)sand CO-PIM-Rs ranged from 500 to 900 m2/g, with adequatemicropore area (438e638 m2/g). The micropore pore volumesranging from 0.22 to 0.44 cm3/g. The difference in specific surfacearea might be due to the structural geometry of the monomersemployed as well as the efficient polycondensation by the incor-poration of highly reactive fluorinated monomers into the reactionmixture, resulting in high-molecular-weight copolymers. Whencomparing the porous properties PIM-Rs and CO-PIM-Rs which arebuilt up from differentmonomers and comonomers, a trend linkingporosity and the comonomer unit can be seen. Typical example ofnitrogen adsorption/desorption isotherm is shown in Fig. 5. Theadsorption isotherms show high uptake at very low relative pres-sure, which is characteristic of a microporous material andisotherms of the prepared polymers exhibit a combination of I andII N2 sorption isotherm with a continuous increase after theadsorption at low relative pressure and a broad hysteresis upondesorption. Suggesting that, a broad hysteresis down to low pres-sure is indicative of trapping effect at cryogenic temperature. Theextremely low relative pressure portion of the adsorption isothermis related to the apparent micropore distribution, which in turn isexpected to have a profound influence on the transport propertiesof the polymer. From the t-plot analysis it was found that all theprepared polymers showed significant amount of microporosity.Surface area for CO-PIM-Rs found to be higher than its related

der at different magnifications.

Table 2Porous properties of PIM-Rs and CO-PIM-Rs and their comparison with PIM-1 and OFP-3.

PIMs SABET (m2/g)a PVmicro (cm3/g) HK median porewidth (Å)

SALAN (m2/g)b

(77 K/87 K)H2 (wt.%)/1.13 bar,(77 K/87 K)

Qst (kJ/mole)

PIM-1 [14b] 860 e 6e7 e 1.04 e

OFP-3 [4] 1159 (1089) 0.69 5e6 934/801 1.56/1.36 7.45PIM-R1 702 (575) 0.28 7.5 432/323 1.01/0.76 7.24PIM-R2 595 (548) 0.25 6.6 343/275 0.79/0.63 7.35PIM-R3 628 (467) 0.23 8.2 467/353 0.97/0.74 7.14PIM-R4 889 (717) 0.37 7.6 531/436 1.18/0.92 6.85PIM-R5 636 (457) 0.23 8.3 422/315 0.86/0.63 6.66PIM-R6 714 (632) 0.32 6.9 453/326 1.03/0.87 7.24PIM-R7 680 (638) 0.33 6.8 440/346 0.92/0.81 7.37CO-PIM-R3 843 (813) 0.41 6.6 502/406 1.08/0.92 6.60CO-PIM-R4 880 (840) 0.42 6.5 545/451 1.26/0.95 7.54CO-PIM-R6 855 (789) 0.44 6.7 541/422 1.21/0.96 8.55CO-PIM-R7 857 (549) 0.37 7.8 553/443 1.17/0.91 7.37

a BET surface area calculated from nitrogen adsorption isotherm. The number in the parenthesis is the micropore surface area calculated using the t-plot analysis. PVmicro isthe micropore volume.

b Surface area calculated from the H2 adsorption isotherm using Langmuir equation at 77 K and 87 K.

Fig. 5. Nitrogen sorption isotherm of CO-PIM-R4 at 77 K and the inset picture ismicropore size distribution as calculated using HK method.

S. Makhseed et al. / Polymer 53 (2012) 2964e29722970

polymer containing the common repeating unit. Analysis of lowrelative pressure region of the N2 adsorption isotherm by theHorvath-Kawazoe (H-K) method assuming a slit-pore geometrygives the apparent micropore distributions is given in the inset ofFig. 5. To inducemicroporosity in a prepared polymer at least one ofthe monomer contain a site of contortion [28], which in many casesrepresented by tetrahedral carbon of a spiral core and moreoverpore distribution with the majority of micropores populated atwidth less than 10 Å. Such narrow pore size distribution is advan-tageous for better interaction with H2 molecules and there byenhancing the storage capacity. The high surface areas of intrinsi-cally porous polymers can be illustrated by its rigid nonlineararchitecture. It is believed that nonlinear polymers can possesslarge amount of interconnected void space. The presence of theseinterconnected voids enhances the intrinsic microporosity of thepolymer even without the network structure. The spiro centrepresent in the 5,50,6,60-tetrahydroxy-3,3,3,30-tetramethyl-1,10-spi-robisindane creates the nonlinearity by propagating the polymerchain in three-dimensional irregular fashion producing a polymerchain that can also prevent the dense packing. The dioxane linkagein conjunction with spiro centre forcing perpendicularly arrangedaromatic rings creates the conditions of supramolecular controlover the molecular structure.

The porous properties of some selected film samples weretested by means of nitrogen adsorption at 77 K. As expected themeasurement were very slow and the film samples showed lowersurface area as compared to its powder form, which may be relatedto the more densely packed structure as confirmed by SEM. Fig. 6shows the N2 adsorption/desorption isotherm of CO-PIM-R4 filmwith a remarkable BET surface area (467 m2/g). The isothermdisplays a considerable uptake at relative low pressure with largerhysteresis due to the swelling/trapping effect [18]. In addition,some film samples show a sudden jump at certain points which canbe explained by the getting pressure effect (see supporting infor-mation). In comparison to powder samples, all the tested filmsamples shows same behaviour in the adsorption isothermwhich istypical for finely microporous media.

3.2. Hydrogen adsorption

The volumetric hydrogen adsorption capacities of all PIMs weremeasured using a ASAP 2020 instrument with 99.999% pure H2over the pressure range 0e1.13 bar. The sorption of hydrogen atcryogenic temperature is also an important tool to identify themicroporosity in polymers. Fig. 7 shows the adsorption/desorptionisotherms at 77 and 87 K for the CO-PIM-R4 with a maximum

hydrogen adsorption capacities up to 1.24 and 0.96 wt% respec-tively at 1.12 bar. The isotherms are fully reversible and exhibita sharp rise at low pressure region which is consistent with thephysisorption of hydrogen on a microporous material. All theprepared PIM-Rs and CO-PIM-Rs show similar behaviour in theirisotherm with a significant uptake at two different temperatures(Table 2). From the isotherms it was found that there was no kinetictrapping of hydrogen in small pores upon desorption. TheHydrogen adsorption capacities by various samples with differentsurface areas usually reveal a linear relationship between BETsurface area and H2 storage capacity at low pressure. For examplethe H2 adsorption capacity of PIM-R2 (0.79 wt.%) and PIM-R4(1.18 wt.%) at 1 bar and 77 K with their BET surface areas 595 and889 m2/g respectively, is consistent with the above hypothesis.Thus from the comparison of various synthesised PIM-Rs andCO-PIM-Rs, the H2 adsorption capacities are largely influenced byBET surface area. Several studies showed that a clear relationshipexists between total surface area and hydrogen adsorption densityof a porous material [29�31]. The hydrogen adsorption at 77 K,physisorption mechanism is dominant and the H2 uptake iscontrolled by the structural features of the adsorbent material.Moreover small micropores can effectively adsorb hydrogen,probably due to its much smaller kinetic diameter compared tobigger gas molecules such as N2 [2b,32]. These ultramicroporesallow the dihydrogenmolecule to interact withmultiple portions of

Fig. 6. Nitrogen sorption isotherm of CO-PIM-R4 film at 77 K.

Fig. 8. Isosteric heats of adsorption for H2 on PIM-R6 and CO-PIM-R6.

S. Makhseed et al. / Polymer 53 (2012) 2964e2972 2971

the double stranded framework and thereby increasing adsorptionpotential and a stronger van der Waals interaction with thehydrogen molecules [33,34]. It is therefore clear that the hydrogenstorage capacity is generally relative to their respective specificsurface areas as well the presence of ultramicroporosity. Thelangmuir model is frequently used to measure specific surface areaof microporous polymers from hydrogen sorption isotherms andthe values ranging upto 550 m2/g as depicted in Table 2. TheLangmuir isotherm theory assumes monolayer coverage of adsor-bate over a homogenous adsorbent surface of equal energy. Theisosteric heat of adsorption, (Qst), for dihydrogen molecules on allsamples was calculated from the adsorption isotherms at 77 and87 K as shown in Fig. 7. The calculated Qst values (6.6e8.5 kJ/mol)are comparable with other porous organic materials such as COFs,PAFs and HCPs. The highest heat of sorption (8.5 kJ/mol) wasobserved for CO-PIM-R6indicating a favourable interactionsbetween dihydrogen and polymer surface. As shown in typicalexample in Fig. 8 the values of Qst decreases rapidly as the functionof degree of coverage. This is attributed to the heterogeneousnature of polymer surface (sorption places are energeticallydifferent) which is available for adsorption [2]. The chemical natureof the accessable surfaces andmorphology of the pores have thus tobe tuned to enable a high adsorption enthalpy. It has been alreadyknown that microstructure had a deep impact on hydrogen uptakewhich is further influenced by the geometry of the building units.

Fig. 7. Hydrogen sorption isotherms of CO-PIM-R4 at 77 K and 87 K.

In a recent study based on FTIR spectroscopy revealed that thedihydrogen molecules adsorb perpendicular to the surface, thepreferential adsorption site is benzene rings [35]. So materialscontaining high electron rich sites were found to increase thehydrogen adsorption enthalpy. The combination of microporosityand chemical functionality of these PIMs are very promising forapplications in the air purification, or separation. As compared toMOFs and similar metal containing porous structures, the conceptof preparing hydro-thermally stable PIMs having organic spacersalong with predefined functionality allows them to extend in manyadsorption applications.

4. Conclusions

In conclusion, a new series of soluble phthalimide based organicpolymers exhibiting intrinsic microporosity were synthesized bythe nucleophilic substitution reaction using various fluoro-monomers and 5,50,6, 60-Tetrahydroxy-3,3,30,30-tetramethylspir-obisindane. The resulting ladder polymers were obtained in highyield. The FTIR, 1H NMR, and elemental analyses indicated that thepolymers appear to have the same proposed structure. Reactionconditions led to high molecular weight polymers which providedsurface areas up to 900 m2 g�1. The pore size distribution analysisshown that pores are narrowly distributed and fine-tuned witha subnanometer dimension, close to the van derWaals diameters ofvarious small gas molecules. The intrinsic microporosity of thesepolymers is supported by good hydrogen uptake at 77 K and 87 K.Isosteric heat of H2 adsorption (Qst) measurement revealed that Qstis not highly sensitive to the structural features and exhibit highisoteric heats of adsorption up to 8.5 kJ/mol.

Acknowledgements

The authors thank Kuwait University for funding (SC03/08) andthe technical support of GFS (GS01/01, GS01/05 and GS01/03) andEMU.

Appendix A. Supporting information

Supplementary data related to this article can be found online atdoi:10.1016/j.polymer.2012.05.001.

S. Makhseed et al. / Polymer 53 (2012) 2964e29722972

References

[1] (a) Lee J-Y, Wood CD, Bradshaw D, Rosseinky MJ, Cooper AI. Chem Commun;2006:2670e2;(b) Wood CD, Tan B, Trewin A, Niu H, Bradshaw D, Rosseinsky MJ, et al. ChemMater 2007;19:2034e48.

[2] (a) Germain J, Hradil J, Frechet JMJ, Svec F. Chem Mater 2006;18:4430e5;(b) Germain J, Svec F, Frechet JMJ. Chem Mater 2008;20:7069e76.

[3] Ghanem BS, McKeown NB, Harris KDM, Pan Z, Budd PM, Butler A, et al. ChemCommun; 2007:67e9.

[4] Makhseed S, Samuel J. Chem Commun; 2008:4342e4.[5] (a) Kuhn P, Antonietti M, Thomas A. Angew Chem Int Ed 2008;47:3450e3;

(b) Kuhn P, Forget A, Su DS, Thomas A, Antonietti MJ. Am Chem Soc 2008;130:13333e7.

[6] Wang Z, Zhang B, Yu H, Sun L, Jiao C, Liu W. Chem Commun 2010;46:7730e2.[7] (a) Jiang JX, Su F, Trewin A, Wood CD, Campbell NL, Niu H, et al. Angew Chem

Int Ed 2007;46:8574e8;(b) Jiang JX, Su F, Trewin A, Wood CD, Campbell NL, Niu H, et al. J Am ChemSoc 2008;130:7710e20.

[8] Dawson R, Laybourn A, Clowes R, Khimyak YZ, Adams DJ, Cooper AI. Macro-molecules 2009;42:8809e16.

[9] Dawson R, Cooper AI, Adams DJ. Prog Polym Sci 2012;37:530e63.[10] Thomas A. Angew Chem Int Ed 2010;49:8328e44.[11] Cooper AI. Adv Mater 2009;21:1291e5.[12] McKeown NB, Budd PM. Macromolecules 2010;43:5163.[13] McKeown NB, Budd PM. Chem Soc Rev 2006;35:675e83.[14] (a) Budd PM, Ghanem BS, Makhseed S, Mckeown NB, Msayib K, Tattershall CE,

et al. Chem Commun 2004;2:230e1;(b) Budd PM, Elabas ES, Ghanem BS, Makhseed S, McKeown NB, Msayib KJ,et al. Adv Mater 2004;16:456e9;(c) McKeown NB, Budd PM, Msayib K, Ghanem BS, Kingston HJ, Tattershall CE,et al. Chem A Eur J 2005;11:2610e20.

[15] Du N, Song J, Robertson GP, Pinnau I, Guiver MD. Macromol Rapid Commun2008;29:783e8.

[16] GhanemBS,McKeownNB,BuddPM,FritschD.Macromolecules2008;41:1640e6.

[17] (a) Du N, Robertson GP, Pinnau I, Guiver MD. Macromolecules 2009;42:6023e30;(b) Du N, Robertson GP, Song J, Pinnau I, Guiver MD. Macromolecules 2009;42:6038e43.

[18] Weber J, Du N, Guiver MD. Macromolecules 2011;44:1763e7.[19] Du N, Park HB, Robertson GP, Dal-Cin MM, Visser T, Scoles L, et al. Nat Mater

2011;10:372e5.[20] Mason CR, Maynard-Atem L, Al-Harbi NM, Budd PM, Bernardo P, Bazzarelli F,

et al. Macromolecules 2011;44:6471e9.[21] Weber J, Su Q, Antonietti M, Thomas A. Macromol Rapid Commun 2007;28:

1871e6.[22] (a) Ritter N, Antonietti M, Thomas A, Senkovska I, Kaskel S, Weber J. Macro-

molecules 2009;42:8017e20;(b) Ritter N, Senkovska I, Kaskel S, Weber J. Macromol Rapid Commun 2011;32:438e43.

[23] Du N, Robertson GP, Pinnau I, Guiver MD. Macromolecules 2010;43:8580e7.

[24] Dawson R, Laybourn A, Khimyak YZ, Adams DJ, Cooper AI. Macromolecules2010;43:8524e30.

[25] Makhseed S, Samuel J, Bumajdad A, Hassan M. J Appl Polym Sci 2008;109:2591e7.

[26] Kricheldorf HR, Schwarz G. Macromol Rapid Commun 2003;24:359e81.[27] Miyatake K, Hlil AR, Hay AS. Macromolecules 2001;34:4288.[28] Carta M, Raftery J, McKeown NB. J Chem Crystallo 2011;41:98e104.[29] Dinca M, Yu AF, Long JR. J Am Chem Soc 2006;128:17153. 17153.[30] Foy AGW, Matzger AJ, Yaghi OM. J Am Chem Soc 2006;128:3494e5.[31] Mckeown NB, Gahnem B, Msayib KJ, Budd PM, Tattershall CE, Mahmood K.

Angew Chem Int Ed 2006;45:1804e7.[32] Weber J, Antonietti M, Thomas A. Macromolecules 2008;41:2880e5.[33] Shiraishi M, Takenobu T, Yamada A, Ata M, Kataura H. Chem Phys Lett 2002;

358:213e8.[34] Yushin G, Dash RK, Jagiello J, Fischer JE, Gogotsi Y. Adv Funct Mater 2006;16:

2288e93.[35] Spoto G, Vitillo JG, Cocina D, Damin A, Bonino F, Zecchina A. Phys Chem Chem

Phys 2007;9:4992e9.